Herpes Simplex and Eosinophilic Oesophagitis: The Chicken or the Egg?

Eosinophilic oesophagitis is being recognised more frequently in developed countries (1). It is more commonly seen in male individuals (2) and presents with symptoms of gastroesophageal reflux, dysphagia, and episodes of food bolus obstruction. Many patients have a positive history of atopy (3). In some if not all patients, the isolated eosinophilic inflammation of the oesophagus is thought to be part of an allergic response to food antigens (4) or aeroallergens (eg, pollen), perhaps triggered by environmental factors that predispose the patient to the development of an eosinophilic response in the presence of such antigens or allergens. One such trigger could be a viral agent such as herpes simplex virus (HSV), a common infection in the community. The following case series describes 3 patients with HSV infection in whom the endoscopic features of eosinophilic oesophagitis developed. The implications of a possible link between the 2 conditions are discussed. Eosinophilic oesophagitis is variously defined in the literature. For the purposes of these case reports, it is defined as more than 20 eosinophils per high-power field (>20/HPF) associated with basal cell hyperplasia, reflecting the recently published Melbourne experience (3,5). CASE REPORT 1 A 14-year-old boy from Melbourne presented with acute severe odynophagia. His medical history included hay fever and frequent episodic asthma, for which he had been taking fluticasone for several years. Gastroscopy was performed, and the macroscopic appearance initially suggested Candida infection with multiple plaques throughout the oesophagus (Fig. 1). However, the microscopic findings were of typical herpesvirus inclusions (Fig. 2A,B). The biopsy specimens were also positive for HSV by polymerase chain reaction and immunoperoxidase stain. He had never had “cold sores” or episodes of gingivostomatitis. Eosinophils and basal cell hyperplasia were not prominent in the biopsy specimens. He was treated with acyclovir and made a prompt recovery. The results of detailed immune function tests were normal, apart from raised immunoglobulin (Ig)E levels (589 kU/L, normal range 0–200). He had skin prick tests earlier in connection with his asthma (results unavailable), and he generally avoided dairy products, consuming soy, if anything.FIG. 1: Multiple white plaques adherent to oesophagus on endoscopic view, an appearance similar to that seen with candidal infection.FIG. 2: A, Low power view of oesophageal tissue containing viral inclusions. There is no basal cell hyperplasia or eosinophilia. (Haematoxylin & eosin stain; original magnification ×40.) B, High-power view of viral inclusion bodies. (Haematoxylin & eosin stain; original magnification ×400.)Repeat endoscopy 2 months later when he had experienced intermittent symptoms showed full-length longitudinal furrowing (Fig. 3) of the oesophagus and several linear ulcers in the lower third. Microscopically, the upper and lower oesophagus showed basal cell hyperplasia and a prominent eosinophilic infiltrate (Fig. 4A, B) suggestive of eosinophilic oesophagitis. There were also changes consistent with gastroesophageal reflux in the biopsy specimens from the lower oesophagus. There was no longer any histological or culture evidence of HSV infection.FIG. 3: Longitudinal furrowing of the oesophagus typical of that seen in eosinophilic oesophagitis.FIG. 4: A, Low-power view of basal cell hyperplasia and eosinophilia. (Haematoxylin & eosin stain; original magnification ×40.) B, High-power view of basal cell hyperplasia and eosinophilia. (Haematoxylin & eosin stain; original magnification ×400.)He was prescribed esomeprazole and underwent repeat endoscopy 2 months later, when he was asymptomatic, although in the interim he had been acutely unwell with chickenpox. This third endoscopy again showed moderate eosinophilic infiltrate and basal cell hyperplasia. Esomeprazole was continued because missed doses resulted in symptoms. When the next endoscopy showed no improvement, oral fluticasone was commenced (250 μg twice daily). After 2 months of fluticasone, repeat endoscopy showed only minimal macroscopic furrowing and only 13 eosinophils per HPF microscopically. Some reflux oesophagitis remained in the lower third, so esomeprazole was continued. Fluticasone was ceased. Although there was no rebound of symptoms after fluticasone was discontinued, missed doses of esomeprazole continued to result in heartburn. After 6 months when he did not take fluticasone, endoscopy was performed and again showed macroscopic features of eosinophilic oesophagitis, and eosinophil counts (upper 26/HPF, lower 30/HPF) and basal cell hyperplasia. Fluticasone was recommenced. CASE REPORT 2 A 16-year-old boy from Melbourne presented acutely with dysphagia and odynophagia. His history included repeated esophageal dilatation of strictures resulting from a caustic injury at 3 years of age. He had also required endoscopic removal of food boluses stuck in his mid-oesophagus at ages 5 and 6. He was subsequently symptom free for the 5 years preceding this presentation. Other history included infrequent episodic asthma and anaphylaxis to legumes. Two of his siblings had food allergies and eczema. It was assumed by the treating surgeon that his presenting dysphagia at age 16 was related to the previous caustic damage, and biopsy of his oesophagus was not performed via therapeutic endoscopy. It is thus not known whether he may have had eosinophilic oesophagitis before this presentation. On referral to a gastroenterologist, gastroscopy showed extensive and severe ulcerative oesophagitis with a heavy infiltrate of polymorphs and changes of gastroesophageal reflux, but no plaques. At histopathological examination, HSV-1 was isolated on serology and culture; however, inclusion bodies were not seen. His symptoms improved after treatment with omeprazole and acyclovir. Extensive testing of his cell- and humoral-mediated immune system was normal, but the results of skin prick tests were strongly positive for peas, lentils, and hazelnuts. He was also found to have serum IgE levels of 737 kU/L (normal 0–200), but he did not have an associated peripheral eosinophilia. Follow-up endoscopy 6 weeks later showed macroscopic changes of longitudinal furrowing of the body of the oesophagus consistent with eosinophilic oesophagitis. The results of biopsy were consistent with this diagnosis, in that he had dense infiltrates of eosinophils in the oesophagus, with 23 eosinophils per HPF in the upper oesophagus, 45 eosinophils per HPF in the mid-oesophagus, and 35 eosinophils per HPF in the lower oesophagus. At this time his HSV culture was negative but with a positive antibody response. He did not have dysphagia at this appointment or at the next appointment 3 months later. The patient then did not return for planned repeat endoscopy 3 months later, and he has been lost to further follow-up. CASE REPORT 3 A 9-year-old boy from urban Rio de Janeiro, with a history of chronic sinus disease and hay fever and a family history of atopy, was brought for treatment after 7 to 10 days of vomiting and fever followed by acute onset of dysphagia. On examination he was found to have tender hepatosplenomegaly and 2 white spots under his tongue. He had never previously experienced dysphagia. Gastroscopy was performed, and ulcers suggestive of herpesvirus or cytomegalovirus (CMV) oesophagitis, bleeding, and necrotic tissue were visualised. There were also some white patches suggestive of Candida infection. Microscopically, the squamous epithelium was detached, and there were prominent groups of leukocytes and also multinucleated giant cells, all suggestive of herpetic oesophagitis. At this time, biopsy revealed fewer than 10 eosinophils per HPF. He was treated with 10 days of intravenous acyclovir and fluconazole (before the histological results disproved suspected Candida infection), and his symptoms resolved over 3 days. His initial serology showed IgG and IgM positivity to HSV; 1 month later the IgM was negative. Subsequent immune testing showed normal levels of immunoglobulins and complement and negativity for human immunodeficiency virus. He had documented previous infection with Epstein-Barr virus and CMV. The results of skin prick testing, which had been performed before the dysphagia for his chronic sinus disease, were positive for dust and dermatophagoids (Acarus). The results of both skin prick and patch testing were negative for cow's milk, soy, meat, fish, and chicken. At follow-up 1 month later, he was found to have a peripheral eosinophilia, and gastroscopy was repeated. The mid-oesophagus contained a marked eosinophilic infiltrate (30/HPF) but showed resolution of the features of oesophagitis. He remained asymptomatic, but repeat endoscopy 2 months later showed persistent findings (eosinophils >30/HPF), and he was prescribed oral fluticasone. He underwent repeat endoscopy 6 months later, after 3 months receiving inhaled fluticasone, with which he had good compliance. Macroscopically, white patches were seen throughout the oesophagus. The results of biopsy were consistent with the diagnosis of eosinophilic oesophagitis. There were infiltrates of eosinophils in the oesophagus, with more than 35 eosinophils per HPF in the upper, mid-, and lower oesophagus. At this time he was asymptomatic. Repeat herpes serology showed positivity for IgM and IgG. He was prescribed acyclovir again but continues to have a small herpetic lesion near his nose. DISCUSSION We report 3 separate patients presenting with odynophagia and erosive oesophagitis associated with acute HSV infection, as evidenced by isolation of the virus and subsequent seroconversion. Subsequently, endoscopic features of eosinophilic oesophagitis developed in all patients, characterised by longitudinal furrowing of the esophageal mucosa and dense eosinophilic infiltrate on histopathology. Although HSV oesophagitis is reported in immunocompetent children, it is uncommon (6–8). Although eosinophilic oesophagitis is becoming more frequently identified, the sequence of events in these 3 patients suggests that the 2 may be linked rather than occurring together by chance alone. We speculate that these findings are not random, and we propose possible mechanisms to explain the phenomenon. One explanation is that HSV infection causes tissue damage and a breakdown of the barrier function of the esophageal mucosa, resulting in a florid immune response with breakdown of immune tolerance and development of hyperreactivity. This mechanism has been postulated in celiac disease and other food allergies, in which it is thought that transient infection may lead to activation of dendritic cells, macrophages, and other immune cells in the lamina propria (9). Alternatively, eosinophilic oesophagitis (perhaps triggered by reflux disease or caustic injury) may increase susceptibility to infection with HSV. A dysregulated T helper 2 response (as can occur in atopic individuals) could lead to a breakdown in barrier function, allowing the virus to establish an infection that is otherwise uncommon in immunocompetent hosts (10). None of the 3 patients had clear evidence for eosinophilic oesophagitis preceding or at the time of presentation with herpes oesophagitis. It is possible but not provable from our cases that the typical endoscopic and histologic features of eosinophilic oesophagitis may have been masked by those of the infection. However, good-size biopsy samples were obtained, and the areas of the biopsy specimens that showed squamous mucosa (not just infected tissue) did not show basal cell hyperplasia or increased eosinophils. The lack of histological evidence of eosinophilic oesophagitis in our cases suggests either that the HSV infection predated or perhaps triggered the development of the eosinophilic oesophagitis, or that eosinophilic oesophagitis existed but was not evident because of the HSV. The role, if any, of the caustic injury in the second case is speculative. To our knowledge there have been no reports of caustic ingestion being associated either with susceptibility to herpes oesophagitis or with the development of eosinophilic oesophagitis. The use of fluticasone as treatment for eosinophilic oesophagitis, symptomatic or otherwise, has yet to emerge as best practice. It can be shown to improve and sometimes to completely reverse the changes of eosinophils oesophagitis, but it is unclear whether resolution of histological changes confers benefit to the patient. Recent reports suggest that the risk-to-benefit ratio leans towards treatment (11,12). Further studies are required. It has been our practice to review patients endoscopically after treatment, particularly in these 3 unusual cases. It is hoped that the report of these 3 cases will bring more cases to light and lead to further discussion. This may help explain the apparent link between herpes oesophagitis and eosinophilic oesophagitis and in turn lead to a greater understanding of the immune mechanisms involved. Colleagues are encouraged to report other cases.